Diseño de un agente inteligente basado en una red neuronal artificial supervisada. Validación en un dominio botánico

Artificial Neural Networks (ANN) were created to simulate the components and functions of the human brain. The properties of ANN enable problems apply in pattern recognition and classification, as they are able to discover no apparent relationships between variables, and both give meaning to data. A...

Descripción completa

Guardado en:
Detalles Bibliográficos
Autores principales: Mariño, Sonia I., Primorac, Carlos R.
Formato: Artículo revista
Lenguaje:Español
Publicado: Escuela de Perfeccionamiento en Investigación Operativa 2021
Materias:
Acceso en línea:https://revistas.unc.edu.ar/index.php/epio/article/view/33195
Aporte de:
Descripción
Sumario:Artificial Neural Networks (ANN) were created to simulate the components and functions of the human brain. The properties of ANN enable problems apply in pattern recognition and classification, as they are able to discover no apparent relationships between variables, and both give meaning to data. An intelligent agent whose performance is monitored element models of RNA is described. Validation was applied to the automatic identification of three plant species of Rollinia: R. salicifolia, R. emarginata and R. rugulosa. The constructed and validated models are presented, and the one selected from the best results obtained by evaluating the Mean Square Error and the percentage of errors, this was incorporated into the intelligent agent. The agent consists of a user interface to ensure its use. The proposal may be adapted to other knowledge’s domains.